- Email: [email protected]
Flavobacterium columnare colony types: Connection to adhesion and virulence?
Microbial Pathogenesis 46 (2009) 21–27
Contents lists available at ScienceDirect
Microbial Pathogenesis journal homepage: www.elsevier.com/locate/micpath
Flavobacterium columnare colony types: Connection to adhesion and virulence? Heidi M.T. Kunttu*, Lotta-Riina Suomalainen, E. Ilmari Jokinen, E. Tellervo Valtonen ¨skyla ¨ , Department of Biological and Environmental Science, P.O. Box 35, FI-40014 Jyva ¨skyla ¨ , Finland University of Jyva
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 May 2008 Received in revised form 3 September 2008 Accepted 4 October 2008 Available online 17 October 2008
Four different colony morphologies were produced by Flavobacterium columnare strains on Shieh agar plate cultures: rhizoid and flat (type 1), non-rhizoid and hard (type 2), round and soft (type 3), and irregularly shaped and soft (type 4). Colonies produced on AO agar differed from these to some extent. The colony types formed on Shieh agar were studied according to molecular characteristics [Amplified Fragment Length Polymorphism (AFLP), Automated Ribosomal Intergenic Spacer Analysis (ARISA), and whole cell protein SDS-PAGE profiles], virulence on rainbow trout fingerlings, and adhesion on polystyrene and fish gills. There were no molecular differences between colony types within one strain. Type 2 was the most adherent on polystyrene, but type 1 was the most virulent. Adhesion of F. columnare strains used in this study was not connected to virulence. From fish infected with colony type 1, three colony types (types 1, 2 and 4) were isolated. Contrary to previous studies, our results suggest that strong adhesion capacity may not be the main virulence factor of F. columnare. Colony morphology change might be caused by phase variation, and different colony types isolated from infected fish may indicate different roles of the colony morphologies in the infection process of columnaris disease. Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Flavobacterium columnare Colony morphology Virulence Adhesion Phase variation
1. Introduction Flavobacterium columnare, the causative agent of columnaris disease, is a freshwater bacterium that can be isolated from natural waters [1,2]. Nowadays, the bacterium is considered as one of the most harmful bacterial fish pathogen at freshwater fish farms worldwide: columnaris outbreaks cause remarkable financial and material losses yearly for the fish farming industry (see Ref. [3]). Development of improved cultivation methods (e.g. Refs. [4,5]) has enabled a routine isolation of F. columnare, facilitating the diagnostics and studying of this pathogen in laboratory conditions. Despite its pathogenicity, virulence mechanisms of F. columnare are largely unknown. It is known, that different genetic groups [6–8] express different degrees of virulence. It has also been shown that the activity of connective tissue degrading enzyme, chondroitin AC lyase [7], and capacity to adhere on gill tissue [9] are related to virulence of F. columnare. Growth characteristics in different culture conditions [10], production of extracellular proteases [11,12] and outer membrane protease genes [13] have been studied, but clear correlation to virulence has not been observed. On the other hand, differences in lipopolysaccharide (LPS) and protein profiles between virulent and avirulent F. columnare strains have been detected [14]. In some pathogenic
* Corresponding author. Tel.: þ358 14 2604228; fax: þ358 14 2604756. E-mail address: hkunttu@jyu.fi (H.M.T. Kunttu). 0882-4010/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.micpath.2008.10.001
bacteria, cell surface components, such as LPS and capsular material, function as virulence factors (e.g. Refs. [15–17]). Change in cell surface components often leads to change in colony morphology of a bacterium (see e.g. Refs. [16–19]). In some human and animal (also fish) pathogens, such as tubercle bacilli [20], Vibrio vulnificus [21], and Mycobacterium avium [22], different colony morphologies produced by one strain can exhibit difference in virulence. Previously, difference in colony morphology has been detected between F. columnare strains [23]. Also non-rhizoid as well as soft and non-adherent colonies have been noticed to appear among rhizoid colonies after subcultivation [24,25]. It has been shown, that spreading or rhizoid colony formation is an indication of gliding motility of Flavobacterium johnsoniae (previously Cytophaga johnsonae) [26] and Flavobacterium psychrophilum (previously Cytophaga psychrophila) [27]. Therefore, it is possible, that the loss of gliding motility appears as non-rhizoid colony morphology also in F. columnare. Different colony morphologies have also been found in Flavobacterium succinicans (previously Cytophaga succinicans), a relative to F. columnare [28]. However, the pathogenesis of colony morphology variants has not been studied in flavobacteria. We found that four different colony morphologies are formed among F. columnare strains on Shieh agar plates in laboratory cultivations, and one strain can form one or two morphology variants. Because of severity of repeated infections at fish farms and lack of knowledge of virulence mechanisms in F. columnare, we consider it important to study further these colony types. This is necessary both for developing exact diagnostic tools for columnaris infections and also to find out whether there is a connection
22
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
between virulence and colony morphology. This is why molecular characteristics and adhesion properties as well as fish mortality caused by different colony morphologies were studied. 2. Results Four different colony types (Fig. 1) were formed on Shieh agar plates in laboratory conditions by the eight F. columnare strains tested (Table 1). Colony type 1 was rhizoid and flat with yellow centre. Colony type 2 was hard, more orange in color, non-rhizoid or only slightly rhizoid, and had irregular edges and convex growth form. Colony type 3 had round edges, and smooth, yellowish appearance. Type 4 colonies were white or light yellow, smooth and spreading on the agar with irregular shape. Originally, F. columnare strains formed colony types 1–4 (Table 1), each strain producing only one colony type. In further plate cultivations other colony types started to form among original types in a following manner: type 1 / type 2, type 2 / type 4, and type 3 / type 4, meaning that among type 1 colonies, type 2 colonies started to appear, etc. These other colony types all formed in the cultures of same age and, once formed, remained the same from generation to generation in laboratory cultivations. There was no colony type change in the opposite direction meaning, that no type 3, 2 or 1 colonies were formed among type 4 colonies, no type 2 or 1 colonies among type 3 colonies, and no type 1 colonies among type 2 colonies. Colony type did not change in broth cultivations. In older cultures (more
than 3 days after plate cultivation), however, on the edges of some types 2 and 3 colonies growth resembling type 4 started to appear. There were no contaminations of other bacteria or between F. columnare strains in the cultures. Colony type 4 existed only in the genetic groups with low or intermediate virulence, whereas colony type 3 existed only in high virulence strain. On AO agar, formation of colony morphologies differed from that on Shieh agar to some extent: The growth form of bacteria was more spreading on AO agar than on Shieh agar. Moreover, the older cultures on AO agar started to get transparent appearance, which did not occur on Shieh agar. The growth form of all bacteria of all genetic groups resembled colony type 4 formed on Shieh agar, but had more spreading colonies with slightly rhizoid edges. The exceptions were genetic groups B and F, which formed types 3 and 4 (formed on Shieh agar) resembling colonies also on AO agar, respectively. Among rhizoid type 4 colonies, also types 1 and 2 (formed on Shieh agar) resembling colonies were formed by bacteria of genetic groups D, E and G. None of the colony types 1–4 formed on AO agar had exactly the same appearance as the corresponding colony types 1–4 formed on Shieh agar, but could be categorized into these groups. However, because of the spreading growth form of bacteria on AO agar, the different colony types grew rather stuck on each other than as separate colonies. On AO agar, the colony types 1, 2 and 4 formed originally on Shieh agar grew as rhizoid type 4 colonies. Types 1 and 2 (formed on Shieh agar) resembling colonies were also detected among D2 (colony type 2 formed by strain D), E1, E2 and G2. There were no differences in ARISA [10], AFLP or whole cell protein SDS-PAGE (Fig. 2) profiles between different colony types formed by one F. columnare strain. However, there were differences in adhesion capacities on polystyrene (Kruskal–Wallis test, overall comparison between strains A–H forming colony types 1–4: c2 ¼ 66,947, df ¼ 12, P < 0.001) (Fig. 3a): D2 and D4, E1 and E2, G2 and G4, as well as H2 and H4 differed significantly from each other. Similarly, on polystyrene, there was a significant difference between adhesion capacities of colony types (one-way ANOVA: F ¼ 36.679, df ¼ 4, P < 0.001) (Fig. 3b); colony type 2 was more adherent than other colony types, but also the adhesion of colony type 1 differed significantly from the other colony types. The growth rate of bacteria did not affect the adhesion capacities. On agar, colony type 2 was the most adherent followed by types 1, 3 and 4. This result was achieved by the experience when handling the colonies with inoculation loop. Adhesion capacities on gill tissue differed significantly between colony types among the strains (one-way ANOVA: F ¼ 3,306, df ¼ 6, P ¼ 0.031) (Fig. 3c). G4 was the most adherent and differed significantly from all the other colony types tested, except from E2. A significant difference between adhesion capacities of E2 and H4 was also detected. In challenge experiment, there was a significant difference between mortalities caused by the colony types (Kaplan–Meier Table 1 Colony morphology types formed on Shieh agar in laboratory conditions by genetically grouped Flavobacterium columnare strains (see Refs. [10] and [7] for more details on the strains).
Fig. 1. Representatives of different colony morphologies (1 ¼ colony type 1, 2 ¼ type 2, 3 ¼ type 3, 4 ¼ type 4) formed by Flavobacterium columnare (see also Table 1) on Shieh agar plate cultivation in laboratory. Type 1 is rhizoid and flat with yellow centre. Type 2 is hard, more orange in color, non-rhizoid or only slightly rhizoid, and has irregular edges and convex growth form. Type 3 has round edges, and smooth, yellowish appearance. Type 4 colonies are white or light yellow, smooth and spreading on the agar with irregular shape.
Genetic group
Virulence of the strain
Original colony type
Colony type formed in subcultivations
A B C D E F G H
Low High High Low High Low Intermediate Low
1 3 1 2 1a 4 2a 2a
4 2a 4a 4a
a Colony type was used in the virulence experiment using rainbow trout (Oncorhynchus mykiss, Walbaum) fingerlings.
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
M
A1
B3
C1 D2 D4
E1
E2
F4
G2 G4
H2
H4
M
a 0,080
250 130
72 55
**
0,070
OD (595 nm)
95
23
0,060
**
**
0,050
**
0,040 0,030 0,020 0,010 0,000 B3 C1 E1 E2
High virulence
36
G2 G4
A1 D2 D4 F4 H2 H4
Intermediate virulence
Low virulence
E. coli
Genetic group and colony type 28
b
0,1 0,09
* **
survival analysis: c2 ¼ 187.483, df ¼ 6, P < 0.001) (Fig. 4). Interestingly, the most polystyrene-adherent colony type 2 did not produce the highest mortality. The cumulative mortality caused by E1 was the highest, 60%, and statistically different from mortalities caused by the other colony types tested (E2, G2, G4, H2 and H4) (Table 2). The symptoms of columnaris disease (skin and/or gill lesions) were detectable in all fish infected with E1, but present approximately only in half of the fish infected with the other colony types. F. columnare could be isolated only from fish infected with colony type E1. Interestingly, from primary cultivations made from these fish, several colony types per one fish could be isolated: in addition to colony type 1, the isolated bacteria were also of colony types 2 and 4. 3. Discussion This is the first time when spontaneously formed colony morphologies among F. columnare strains are studied to find connection between colony types, adherence and virulence. Colony types among each F. columnare strain showed no genomic differences in ARISA, AFLP, or whole cell protein SDS-PAGE profiles. The four colony types formed on Shieh agar by F. columnare strains differed significantly in their adhesion capacities on polystyrene and agar; type 2 was the most adherent followed by types 1, 3 and 4. Virulence, however, was not related to adherence capacity: strongly polystyrene- and agar-adherent type 2 bacteria (E2, G2 and H2) were not more virulent than rhizoid type 1 (E1) or poorly adherent type 4 (G4 and H4) bacteria. F. columnare could be isolated only from fish dead in E1 infections, which may be due to difficulties in isolation of F. columnare at preclinical stages of infection, or possible presence of competitive bacteria [29]. Our results suggest that strong adhesion is not essential for virulence of F. columnare. In addition, adhesion to gill tissue was not related either to virulence or adherence on polystyrene, indicating that same molecules are not used to attach on different surfaces, and that virulence mechanisms other than adhesion may be important in pathogenesis of F. columnare. Formation of different colony types occurred also when F. columnare was grown on AO agar, but colonies were not exactly similar to and did not grow as clear separate colonies as the colonies formed on Shieh agar. Differential colony formation on Shieh and AO agar may have been caused by different nutrition composition between these two media. Shieh medium is
*
0,07
* **
0,06
* **
* **
0,05 0,04 0,03 0,02 0,01 0
1
2
3
4
Colony type
c
500
*
450
*
400
CFU ml-1 x 106
Fig. 2. Whole cell protein SDS-PAGE profiles of Flavobacterium columnare strains A–H forming colony types 1–4 (see Table 1). In the first and the last lane are the molecular mass markers (kDa).
OD (595 nm)
0,08
**
* *
350
*
300 250 200 150 100 50 0 C1
E1
E2
High virulence
G2
G4
Intermediate virulence
H2
H4
Low virulence
Genetic group and colony type Fig. 3. (a). Adherence (mean SE) on polystyrene of Flavobacterium columnare strains A–H forming colony types 1–4 (see Table 1), and Escherichia coli. Statistical differences (Kruskal–Wallis test, pairwise comparisons) between different colony types formed by one strain: * ¼ P < 0.05, ** ¼ P < 0.01, *** ¼ P < 0.001. (b). Adherence (mean SE) on polystyrene of Flavobacterium columnare strains (A–H) representing the same colony morphology type (1–4) (see Table 1). Statistical differences (one-way ANOVA, pairwise comparisons, Post Hoc LSD test) between different colony types: * ¼ P < 0.05, ** ¼ P < 0.01, *** ¼ P < 0.001. (c). Adherence (mean SE) on gill tissue (rainbow trout (Oncorhynchus mykiss, Walbaum) of Flavobacterium columnare strains C, E, G and H forming colony types 1–4 (see Table 1). Statistical differences (one-way ANOVA, pairwise comparisons, Post Hoc LSD test) between the strains: * ¼ P < 0.05, ** ¼ P < 0.01, *** ¼ P < 0.001).
specific for F. columnare [5] and might thus allow broader variation of colonial morphology than AO medium. Change in colony morphology is not a rare phenomenon; it occurs in many bacterial species pathogenic to humans and animals. For example, colony morphology variation between translucent and opaque colony types is observed both in Vibrio parahaemolyticus [30] and V. vulnificus [21], whereas M. avium [22]
24
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
100 E1 E2 G2 G4 H2 H4 Control
Cumulative mortality (%)
90 80 70 60 50 40 30 20 10 0 0
24
48
72
96
120
144
168
192
Time (hours after exposure) Fig. 4. Average cumulative mortality of rainbow trout (Oncorhynchus mykiss, Walbaum) fingerlings challenged with equal number of bacteria representing different morphology types (types 1, 2 and 4, see Table 1) formed by Flavobacterium columnare genetic groups E (high virulence), G (intermediate virulence) and H (low virulence).
represents smooth and rough colony types. Colony morphology change is sometimes caused by a phenomenon known as phase variation (reviewed in Refs. [31,32]). Phase variation is often thought to be a random event occurring in a bacterial population as a consequence of heritable genomic changes which can be seen as subpopulations of different phenotypes [33]. Most of the changes are caused by more or less radical genomic rearrangements, like DNA inversions or recombinational deletions, but some minor DNA modifications, like differential DNA methylation, are also known to lead to phase variation. However, growing amount of evidence suggests that phase variation is not always a random event, but a response to an environmental change and caused by change in gene expression [31]. Phase variation allows bacteria to adapt the surrounding environment and is caused by a reversible on/off switch in the expression of one or more genes [32]. Switch can be triggered by intra- or extracellular activators or inhibitors, such as temperature [34], or signaling molecules received from the environment [32]. These molecules can be related to e.g. bacterial quorum sensing. Quorum sensing is a chemical signaling system which is used by bacterial populations to control variety of functions via regulation of gene expression [35]. Besides chemicals produced by bacteria, molecules received from a host can also affect behavior of bacterial populations [36,37]. Production of non-typical colony morphologies after subcultivation has been previously detected also in F. columnare [25]. In another study, non-adherent F. columnare colony variant was obtained by b-lactam selection [24]. In these studies, no genetic methods were used to find out whether different colony formation is caused by changes at the DNA level. Studying the properties of different colony types can provide information about virulence and
Table 2 Difference between colony morphology types 1, 2 and 4 formed by Flavobacterium columnare genetic groups E, G and H (see also Table 1) in virulence on rainbow trout (Oncorhynchus mykiss, Walbaum) fingerlings (Kaplan–Meier survival analysis, pairwise comparisons, Mantel Cox). Genetic group and colony type
E2 G2 G4 H2 H4 a
Difference in virulence: P-values E1
E2
G2
G4
H2
0.000a 0.000a 0.000a 0.000a 0.000a
0.587 0.787 0.002a 0.070
0.778 0.007a 0.197
0.004a 0.119
0.081
Statistically significant difference in virulence.
transmission strategies of F. columnare. It is possible that change in colony type of F. columnare is caused by phase variation affecting the expression or modification of some bacterial surface molecule. However, because no genomic changes were detectable by ARISA, AFLP or whole cell protein SDS-PAGE, the changes may be only minor or may take place at the gene expression level. In human pathogens E. coli and Salmonella enterica serotype Typhimurium phase variation occurs as a consequence of different DNA methylation, which does not lead to change in DNA sequence [32]. In V. parahaemolyticus genetic alterations in cell type controlling transcription factor coding opaR gene may lead to formation of different colony types [30]. Although phase variation is usually considered as a reversible phenomenon, also irreversible changes in bacterium’s phenotype can be caused. For example, irreversible phase variation occurs in a human pathogen V. vulnificus [18]. Also in the present study, the change in colony type was found to be irreversible. Possible phase variation occurring in F. columnare may be a consequence of a change in gene expression and a response to a change in growth environment. It is known, for example, that phase variation occurs during starvation in V. vulnificus [38] and Neisseria gonorrhoeae [39]. The effect of change in nutritional conditions of the growth environment on F. columnare phenotype could be seen also in this study as difference between colony morphology formation on Shieh and AO agars. In addition to changes in growth environment, also quorum sensing has a central role in phase variation and biofilm development in some bacteria, for example in V. parahaemolyticus [19]. Biofilm production properties are needed in the beginning of V. cholerae infection, but suppressed later by quorum sensing [40]. Quorum sensing also controls virulence gene expression in V. cholerae [41]. Quorum sensing has not been studied in F. columnare, and in F. psychrophilum no evidence of quorum sensing was detected, when acylated homoserine lactone (AHL)-based quorum sensing signaling system was studied [42]. However, among AHLbased signaling also several other quorum sensing systems are known and the possible existence of quorum sensing system in F. columnare cannot be excluded. Isolation of colony types 1, 2 and 4 from fish infected with E1 indicates that different colony types may have different roles in the infection cycle, but whether they are formed by phase variation as a response to nutritional factors or quorum sensing can only be speculated and needs more studies. However, it is known that, especially in an animal host, the environment can drive the selection against a particular phenotype of a bacterium, which can result in a different phenotype for cells cultured in the laboratory and those isolated from the host [32]. Indeed, phase variation is considered as one mechanism of pathogenic bacteria to avoid immune defense of a host [31]. Our results show, that there is no clear relation between adhesion capacity and virulence of F. columnare strains tested in this study. We have previously tested adhesion capacity of F. columnare also on crude mucus-coated slides [7], and similarly no correlation to virulence was detected. These results contradict the previous study by Bader et al. [24]. In their immersion challenge experiment the smooth colony forming, less virulent, F. columnare mutant was significantly less adherent both on skin and gill tissue of channel catfish (Ictalurus punctatus) than wild type strain, which formed hard colonies with finger-like projections and which may have resembled our colony type 1. However, in their study, Bader et al. did not detect colony morphologies which would have been adherent and hard with no finger-like projections, i.e. similar to our colony type 2. This may explain the difference in results compared to our study: the virulent strain in that study resembling our high virulence colony type 1 was also the most adherent and Bader et al. did not detect highly adherent type 2 colonies, which produced less mortality in our study. There is also a contradiction with the study by Decostere et al. [9], in which study one highly virulent
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
F. columnare strain was more adherent on the gill tissue of black mollies (Poecilia sphenops) than a less virulent strain. Decostere et al. [9] did not provide any further information on the bacterial colonies, so comparison between their and our study is more difficult to make. In our study, the results from eight genetically different strains, representing high, intermediate and low virulence strains, suggest that adherence capacity is not the main factor defining virulence in F. columnare. Further, the adhesion factors of F. columnare in polystyrene adhesion are different from those in gill tissue adhesion. In summary, we suggest, that strong adhesion capacity is not the defining virulence factor of F. columnare. Thus, some other virulence factors, such as chondroitin AC lyase activity [7] or features not yet found, are needed for pathogenesis in this bacterium. In addition, rhizoid colony morphology may also be related to virulence. This has already been shown in F. psychrophilum, when loss of gliding motility led to decreased virulence [43]. Moreover, in addition to F. psychrophilum [27], relation between rhizoid colony formation and gliding motility has been shown on F. johnsoniae [26,44]. However, our observation is based only on one F. columnare strain (E1), but the virulence experiment suggests similar gliding motility–virulence relationship as found in F. psychrophilum. To confirm this, more studies on the effects of colony morphology on bacterial virulence are needed. 4. Materials and methods 4.1. Bacterial strains and isolation of colony types F. columnare type strain F (NCIMB 2248T) and Finnish strains A–E and G–H isolated from diseased fish at different fish farms (see Ref. [10]) and exhibiting different degrees of virulence [7], were stored in 80 C in 500 ml stocks containing 10% fetal calf serum (FCS, Gibco, BRL Co.) and 10% glycerol, thawed and inoculated to 10 ml of Shieh broth [5] (broth I). Cultures were grown at room temperature (RT) under constant shaking (120 rpm) to late log phase (23 h), after which the cultures were plated on Shieh agar (plate I). Plate cultures were grown for 3 days, different colony types isolated with a loop and inoculated to 10 ml of Shieh broth (broth II). Cultures were plated as previously (plate II), grown for 4 days, different colony types isolated and grown in Shieh broth (broth III) and plated (plate III) as previously. From broth III (broth II in case of strain F), stocks were made and stored in 80 C. Plate I cultures were photographed using stereo microscope on days 2, 3 and 6, plate II cultures on days 2 and 4, and plate III cultures on days 2, 3, 4, 7, 9, 10, 17 and 22 after cultivation. For strain F only plate I and II cultures were made, and they were photographed on days 4, 7, 9, 10, 17 and 22 after cultivation. Several consecutive cultivations were made in order to check whether the colony types retained their morphology. For subsequent analyses, bacterial stocks made from broth III (broth II in case of strain F) were thawed and grown in 10 ml of Shieh broth to late log phase. Colony formation on another medium, AO broth and AO agar [45], was also tested. Experiment on AO medium was conducted as follows: F. columnare strains mentioned above were thawed, inoculated to 10 ml of AO broth and grown at RT under constant shaking (100 rpm) for 42 h, after which the cultures were plated on AO agar. Agar cultures were examined for colony formation using stereo microscope on days 2, 5, 7, 19 and 21 after cultivation. Also the colony types originally formed on Shieh agar (see Table 1) were cultivated in AO broth and on AO agar plates and plate cultures were examined as above. 4.2. Adhesion on polystyrene Optical densities (OD) of bacterial cultivations at 595 nm were measured with micro plate reader (Multiskan Plus, Labsystems,
25
Vantaa, Finland) and set to OD 0.260 in 200-ml volume, corresponding to 3.0 108 colony forming units (CFU) ml1 of bacteria. For adhesion, 100 ml of each colony type culture and Escherichia coli strain JM109 as a control was inoculated onto 96-well polystyrene plates (Immuno Plate, Maxisorp, Nunc Co., Denmark) in six replicates. Another plate with four replicate wells of each colony type was used to detect bacterial growth rate during adhesion experiment. Bacteria were allowed to adhere on the plate for 2 h at RT. Relative adherence of bacterial cells was determined by the method described by A´lvarez et al. [43]. Adhesion capacities were determined from OD values at 595 nm after crystal violet staining of adhered bacteria. ODs of growth controls at 595 nm were measured in the beginning and after the adhesion incubation. 4.3. Adhesion on gill tissue The outermost gill arches (mean weight 30 mg) were taken from 12 rainbow trout fry (mean weight 23 g) and exposed to broth cultures of four F. columnare strains (C, E, G and H) representing colony types 1, 2 and 4 (C1, E1, E2, G2, G4, H2 and H4, see Table 1). ODs of bacterial cultivations were set to 0.119 (corresponding to 1.29 108 CFU ml1). One gill arch per treatment was placed on a petri dish and 15 ml of bacterial culture was added. Shieh broth was used as a negative control. For each bacterium, three replicate dishes were made. Bacteria were allowed to adhere on gill tissue for 2 h at RT under constant shaking (50 rpm). After adhesion, gill arches were rinsed with 10 ml of sterile distilled water and homogenized in eppendorf tubes containing 300 ml phosphatebuffered saline (PBS). From homogenate 1:10 serial dilutions were made in PBS, and 100 ml of pure homogenate and each dilution were plated on Shieh agar containing tobramycin, 1 mg ml1 [4]. Plates were incubated at RT for 2 days after which plate counts were made and the adhesion capacities of the colony types determined from CFU ml1 values. 4.4. ARISA (automated ribosomal intergenic spacer analysis) Genomic DNA was extracted from colony types grown in Shieh broth using microbial DNA isolation kit (MOBIO). ARISA was conducted as previously [10] with following modifications: reaction mixtures contained 1.0 mM of each primer, 0.2 mM of dNTPs, 1 Biotools DNA polymerase buffer (Biotools), and 0.3 U of Biotools DNA polymerase (Biotools). The total reaction volume in each PCR tube was 30 ml plus 1.0 ml of template DNA, and the temperature of an initial denaturation step 95 C. 4.5. AFLP (amplified fragment length polymorphism) Genomic DNA of colony types was extracted as described above and subjected to AFLP based on Vos et al. [46]. Genomic DNA was digested successively with two restriction enzymes, EcoRI and MseI. Reaction solution (total volume 5.1 ml per sample), containing 2.0 U of EcoRI (Fermentas), 2.0 U of MseI (Fermentas), 0.5 ml of Rþ buffer (Fermentas) and 4.2 ml DNA (ca. 20 ng), was incubated first at 37 C for 1 h (for EcoRI digestion) and then at 65 C for 2 h (for MseI digestion). By heating at 95 C for 5 min and then cooling slowly, double stranded adapters for restriction sites were hybridized from single stranded EcoRI forward and EcoRI reverse (5.0 mM suspension), and MseI ad1 and MseI ad2 (50 mM suspension) adapters. These two stranded adapters were ligated at the ends of restriction fragments by incubating ligation solution (total volume 10 ml per sample), including 1.0 U of T4 DNA ligase (Fermentas), 1 T4 ligation buffer (Fermentas), 0.1 mM of EcoRI adapter, 1.0 mM of MseI adapter, 3.4 ml of sterile distilled water and 5.0 ml of digestion product, at RT for 1 h. The ligation products were used as templates in PCR, in which adapter specific primers were used: one reaction
26
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
was done using primer pair MseI-0/EcorRI-A-IRD700 and another using MseI-0/EcorRI-ACT-IRD800, in which A and ACT stand for additional selective nucleotides and 0 for no selective nucleotide. Twenty microliters of PCR reaction mixture, containing 0.25 mM of each primer, 1.0 mM of MgCl2 (Fermentas), 0.08 mM of dNTPs, 0.2 U of Taq DNA polymerase (Fermentas), 1 Tag DNA polymerase buffer [(NH4)2SO4 added] (Fermentas), 11.2 ml of sterile distilled water, and 3.0 ml of template, was made per each sample. PCR was run according to following program: initial denaturation step at 94 C for 2 min, 13 cycles of amplification I (94 C for 30 s, 65 C for 30 s, 72 C for 2 min), 24 cycles of amplification II (94 C for 30 s, 56 C for 30 s, 72 C for 2 min), and final elongation step at 72 C for 10 min. The AFLP was run with a LI-COR 4200 automatic sequencer (LI-COR BioTech) in 6% Long Ranger denaturing polyacrylamide gel (FMC Bioproducts), and fragment length data was analyzed using IrfanView 3.97 free software.
4.6. SDS-PAGE (sodium dodecyl sulphate-polyacrylamide gel electrophoresis) of whole cell proteins Whole cell proteins were extracted from broth cultures of colony types according to protocol by Tan et al. [47] with slight modifications. Bacterial culture, 100 ml, was centrifuged at 4332 g for 10 min. The pellet was washed once with 50 ml of 10 mM Tris-HCl buffer, pH 7.6, and resuspended in 20 ml of the buffer. The extract was sonicated on ice (Branson Digital SonifierÒ, Branson Ultrasonic Corporation, Danburry, Connecticut, USA) for 1 min (1 s pulse with 2 s interval), and 500 ml of sonicate was suspended in an equal volume of 2 electrophoresis treatment buffer and stored in 20 C until analyzed. Prior electrophoresis, the samples were heated at 100 C for 2 min and placed on ice. Discontinuous polyacrylamide gel, 16 cm long and 1.5 mm thick, was used. Samples (50 ml) were pipetted into wells, and electrophoresis was run first with 120 V for 1 h, then with 180 V for 4 h, and finally with 200 V for 1 h. The protein patterns were visualized by staining with PageBlueÔ Protein Staining Solution (Fermentas).
4.7. Challenge experiment Three F. columnare strains (E, G and H) representing colony types 1, 2 and 4 (E1, E2, G2, G4, H2 and H4, see Table 1), were tested for their virulence in rainbow trout (Oncorhynchus mykiss, Walbaum) fingerlings. Strain E has previously been defined as high virulence, G as intermediate virulence, and H as low virulence strain [7]. These strains were chosen because in the present study it was not possible to carry out an experimental set up with all colony types formed by all genetic groups. Fish, (mean weight 6.0 g) originated from a stock of a fish farm in Central Finland, and were acclimated in laboratory conditions for 3 weeks before the experiment. Bacterial infection of fish was carried out under constant aeration in 10-l plastic aquaria, 25 fish in each aquarium, with well water containing 6.5 106 CFU ml1 of bacteria. Each bacterial infection was done in triplicate, a total of 75 fish per treatment. Fish treated with Shieh broth without bacterial inoculate served as negative control. The water temperature was 25 C and infection time 2 h. After infection, the fish were dipped in pure well water and transferred at random into clean, aerated, 10-l plastic flow troughaquaria (T ¼ 25 C). Fish were monitored for 8 days by 12-h intervals for signs of disease and mortality. Bacterial cultivations from skin and gill lesions of up to ten dead fish per treatment were made on Shieh agar. At the end of the experiment cultivations were made from the trunk kidney of four fish per treatment. Cumulative mortality of fish was used as an indicator of bacterial virulence.
4.8. Statistics Differences between adhesion capacities of F. columnare strains forming colony types 1–4 were analyzed using SPSS 14.0 software. In the polystyrene adhesion experiment, comparisons between OD values of each strain and their colony types were made using nonparametric Kruskal–Wallis test. Means of OD values of each colony type were calculated and compared using one-way ANOVA. CFU counts of gill tissue adhesion were analyzed using one-way ANOVA. Data from the bacterial challenge experiment was analyzed using Kaplan–Meier survival analysis. P-values below 0.05 were considered significant.
Acknowledgements The study was supported by a grant from Academy of Finland (# 110754). The authors would like to thank Ellen and Artturi Nyysso¨nen foundation for funding H. M. T. Kunttu and Graduate School of Biological Interactions for other support. We thank also F. Teixeira and M. Martins for technical help and would also acknowledge Dr. P. Rintama¨ki (University of Oulu) and Dr. V. Hirvela¨-Koski (Finnish Food Safety Authority, Oulu) for kindly providing some of the bacterial isolates used in this study.
References [1] Revetta RP, Rodgers MR, Kinkle BK. Isolation and identification of freshwater bacteria antagonistic to Giardia intestinalis cysts. J Water Health 2005;3:83–8. [2] Rickard AH, McBain AJ, Ledder RG, Handley PS, Gilbert P. Coaggregation between freshwater bacteria within biofilm and planktonic communities. FEMS Microbiol Lett 2003;220:133–40. [3] Wagner BA, Wise DJ, Khoo LH, Terhune JS. The epidemiology of bacterial diseases in food-size channel catfish. J Aquat Anim Health 2002;14:263–72. [4] Decostere A, Haesebrouck F, Devriese LA. Shieh medium supplemented with tobramycin for selective isolation of Flavobacterium columnare (Flexibacter columnaris) from diseased fish. J Clin Microbiol 1997;35:322–4. [5] Shieh HS. Studies on the nutrition of a fish pathogen, Flexibacter columnaris. Microbios Lett 1980;13:129–33. [6] Shoemaker CA, Olivares-Fuster O, Arias CR, Klesius PH. Flavobacterium columnare genomovar influences mortality in channel catfish (Ictalurus punctatus). Vet Microbiol 2007;127:353–9. [7] Suomalainen L-R, Tiirola M, Valtonen ET. Chondroitin AC lyase activity is related to virulence of fish pathogenic Flavobacterium columnare. J Fish Dis 2006;29:757–63. [8] Thomas-Jinu S, Goodwin AE. Morphological and genetic characteristics of Flavobacterium columnare isolates: correlations with virulence in fish. J Fish Dis 2004;27:29–36. [9] Decostere A, Haesebrouck F, Charlier G, Ducatelle R. The association of Flavobacterium columnare strains of high and low virulence with gill tissue of black mollies (Poecilia sphenops). Vet Microbiol 1999;67:287–98. [10] Suomalainen L-R, Kunttu H, Valtonen ET, Hirvela¨-Koski V, Tiirola M. Molecular diversity and growth features of Flavobacterium columnare strains isolated in Finland. Dis Aquat Org 2006;70:55–61. [11] Bertolini JM, Rohovec JS. Electrophoretic detection of proteases from different Flexibacter columnaris strains and assessment of their variability. Dis Aquat Org 1992;12:121–8. [12] Newton JC, Wood TM, Hartley MM. Isolation and partial characterization of extracellular proteases produced by isolates of Flavobacterium columnare derived from channel catfish. J Aquat Anim Health 1997;9:75–85. [13] Xie HX, Nie P, Sun BJ. Characterization of two membrane-associated protease genes obtained from screening out-membrane protein genes of Flavobacterium columnare G4. J Fish Dis 2004;27:719–29. [14] Zhang Y, Arias CR, Shoemaker CA, Klesius PH. Comparison of lipopolysaccharide and protein profiles between Flavobacterium columnare strains from different genomovars. J Fish Dis 2006;29:657–63. [15] Amaro C, Fouz B, Biosca EG, Marco-Noales E, Collado R. The lipopolysaccharide O side chain of Vibrio vulnificus serogroup E is a virulence determinant for eels. Infect Immun 1997;65:2475–9. [16] Arias CR, Shoemaker CA, Evans JJ, Klesius PH. A comparative study of Edwardsiella ictaluri parent (EILO) and E. ictaluri rifampicin-mutant (RE-33) isolates using lipopolysaccharides, outer membrane proteins, fatty acids, Biolog, API 20E and genomic analyses. J Fish Dis 2003;26:415–21. [17] Yoshida S, Ogawa M, Mizuguchi Y. Relation of capsular materials and colony opacity to virulence of Vibrio vulnificus. Infect Immun 1985;47:446–51. [18] Hilton T, Rosche T, Froelich B, Smith B, Oliver J. Capsular polysaccharide phase variation in Vibrio vulnificus. Appl Environ Microbiol 2006;72:6986–93.
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27 [19] Enos-Berlage JL, Guvener ZT, Keenan CE, McCarter LL. Genetic determinants of biofilm development of opaque and translucent Vibrio parahaemolyticus. Mol Microbiol 2005;55:1160–82. [20] Smithburn KC. The colony morphology of tubercle bacilli. III. The relation between virulence and colony form. J Exp Med 1935;62:645–63. [21] Simpson LM, White VK, Zane SF, Oliver JD. Correlation between virulence and colony morphology in Vibrio vulnificus. Infect Immun 1987;55:269–72. [22] Kansal RG, Gomez-Flores R, Mehta RT. Change in colony morphology influences the virulence as well as the biochemical properties of the Mycobacterium avium complex. Microb Pathog 1998;25:203–14. [23] Song YL, Fryer JL, Rohovec JS. Comparison of gliding bacteria isolated from fish in North America and other areas of the Pacific Rim. Fish Pathol 1988;23:197–202. [24] Bader JA, Shoemaker CA, Klesius PH. Production, characterization and evaluation of virulence of an adhesion defective mutant of Flavobacterium columnare produced by b-lactam selection. Lett Appl Microbiol 2005;40:123–7. [25] Bernardet JF.´Flexibacter columnaris´: first description in France and comparison with bacterial strains from other origins. Dis Aquat Org 1989;6:37–44. [26] Chang LE, Pate JL, Betzig RJ. Isolation and characterization of nonspreading mutants of the gliding bacterium Cytophaga johnsonae. J Bacteriol 1984;159:26–35. [27] Bernardet JF, Kerouault B. Phenotypic and genomic studies of ‘‘Cytophaga psychrophila’’ isolated from diseased rainbow trout (Oncorhynchus mykiss) in France. Appl Environ Microbiol 1989l;55:1796–800. [28] Anderson RL, Ordal EJ. Cytophaga succinicans sp. n., a factaltatively anaerobic, aquatic myxobacterium. J Bacteriol 1961;81:130–8. [29] Suomainen L-R, Tiirola MA, Valtonen ET. Effect of Pseudomonas sp. MT5 baths on Flavobacterium columnare infection of rainbow trout and on microbial diversity on fish skin and gills. Dis Aquat Org 2005;63:61–8. [30] McCarter LL. OpaR, a homolog of Vibrio harveyi LuxR, controls opacity of Vibrio parahaemolyticus. J Bacteriol 1998;180:3166–73. [31] van der Woude MW. Re-examining the role and random nature of phase variation. FEMS Microbiol Lett 2006;254:190–7. [32] van der Woude MW, Baumler AJ. Phase and antigenic variation in bacteria. Clin Microbiol Rev 2004;17:581–611. [33] Henderson IR, Owen P, Nataro JP. Molecular switches – the ON and OFF of bacterial phase variation. Mol Microbiol 1999;33:919–32.
27
[34] Ratiner YA. Temperature-dependent flagellar antigen phase variation in Escherichia coli. Res Microbiol 1999;150:457–63. [35] Bassler BL. How bacteria talk to each other: regulation of gene expression by quorum sensing. Curr Opin Microbiol 1999;2:582–7. [36] Daniels R, Vanderleyden J, Michiels J. Quorum sensing and swarming migration in bacteria. FEMS Microbiol Rev 2004;28:261–89. [37] Gonzalez JE, Keshavan ND. Messing with bacterial quorum sensing. Microbiol Mol Biol Rev 2006;70:859–75. [38] Chatzidaki-Livanis M, Jones MK, Wright AC. Genetic variation in the Vibrio vulnificus group 1 capsular polysaccharide operon. J Bacteriol 2006; 188:1987–98. [39] Stohl AE, Seifert HS. The recX gene potentiates homologous recombination in Neisseria gonorrhoeae. Mol Microbiol 2001;40:1301–10. [40] Higgins DA, Pomianek ME, Kraml CM, Taylor RK, Semmelhack MF, Bassler BL. The major Vibrio cholerae autoinducer and its role in virulence factor production. Nature 2007;450:883–6. [41] Zhu J, Miller MB, Vance RE, Dziejman M, Bassler BL, Mekalanos JJ. Quorumsensing regulators control virulence gene expression in Vibrio cholerae. Proc Natl Acad Sci U S A 2002;99:3129–34. [42] Bruhn JB, Dalsgaard I, Nielsen KF, Buchholtz C, Larsen JL, Gram L. Quorum sensing signal molecules (acylated homoserine lactones) in gram-negative fish pathogenic bacteria. Dis Aquat Organ 2005;65:43–52. [43] Alvarez B, Secades P, Prieto M, McBride MJ, Guijarro JA. A mutation in Flavobacterium psychrophilum tlpB inhibits gliding motility and induces biofilm formation. Appl Environ Microbiol 2006;72:4044–53. [44] Hunnicutt DW, McBride MJ. Cloning and characterization of the Flavobacterium johnsoniae gliding motility genes gldD and gldE. J Bacteriol 2001;183:4167–75. [45] Anacker RL, Ordal EJ. Study of a bacteriophage infecting myxobacterium Chondrococcus columnaris. J Bacteriol 1955;70:738–41. [46] Vos P, Hogers R, Bleeker M, Reijans M, Van De Lee T, Hornes M, et al. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res 1995; 23:4407–14. [47] Tan ZY, Xu XD, Wang ET, Gao JL, Martinez-Romero E, Chen WX. Phylogenetic and genetic relationships of Mesorhizobium tianshanense and related rhizobia. Int J Syst Bacteriol 1997;47:874–9.
Contents lists available at ScienceDirect
Microbial Pathogenesis journal homepage: www.elsevier.com/locate/micpath
Flavobacterium columnare colony types: Connection to adhesion and virulence? Heidi M.T. Kunttu*, Lotta-Riina Suomalainen, E. Ilmari Jokinen, E. Tellervo Valtonen ¨skyla ¨ , Department of Biological and Environmental Science, P.O. Box 35, FI-40014 Jyva ¨skyla ¨ , Finland University of Jyva
a r t i c l e i n f o
a b s t r a c t
Article history: Received 21 May 2008 Received in revised form 3 September 2008 Accepted 4 October 2008 Available online 17 October 2008
Four different colony morphologies were produced by Flavobacterium columnare strains on Shieh agar plate cultures: rhizoid and flat (type 1), non-rhizoid and hard (type 2), round and soft (type 3), and irregularly shaped and soft (type 4). Colonies produced on AO agar differed from these to some extent. The colony types formed on Shieh agar were studied according to molecular characteristics [Amplified Fragment Length Polymorphism (AFLP), Automated Ribosomal Intergenic Spacer Analysis (ARISA), and whole cell protein SDS-PAGE profiles], virulence on rainbow trout fingerlings, and adhesion on polystyrene and fish gills. There were no molecular differences between colony types within one strain. Type 2 was the most adherent on polystyrene, but type 1 was the most virulent. Adhesion of F. columnare strains used in this study was not connected to virulence. From fish infected with colony type 1, three colony types (types 1, 2 and 4) were isolated. Contrary to previous studies, our results suggest that strong adhesion capacity may not be the main virulence factor of F. columnare. Colony morphology change might be caused by phase variation, and different colony types isolated from infected fish may indicate different roles of the colony morphologies in the infection process of columnaris disease. Ó 2008 Elsevier Ltd. All rights reserved.
Keywords: Flavobacterium columnare Colony morphology Virulence Adhesion Phase variation
1. Introduction Flavobacterium columnare, the causative agent of columnaris disease, is a freshwater bacterium that can be isolated from natural waters [1,2]. Nowadays, the bacterium is considered as one of the most harmful bacterial fish pathogen at freshwater fish farms worldwide: columnaris outbreaks cause remarkable financial and material losses yearly for the fish farming industry (see Ref. [3]). Development of improved cultivation methods (e.g. Refs. [4,5]) has enabled a routine isolation of F. columnare, facilitating the diagnostics and studying of this pathogen in laboratory conditions. Despite its pathogenicity, virulence mechanisms of F. columnare are largely unknown. It is known, that different genetic groups [6–8] express different degrees of virulence. It has also been shown that the activity of connective tissue degrading enzyme, chondroitin AC lyase [7], and capacity to adhere on gill tissue [9] are related to virulence of F. columnare. Growth characteristics in different culture conditions [10], production of extracellular proteases [11,12] and outer membrane protease genes [13] have been studied, but clear correlation to virulence has not been observed. On the other hand, differences in lipopolysaccharide (LPS) and protein profiles between virulent and avirulent F. columnare strains have been detected [14]. In some pathogenic
* Corresponding author. Tel.: þ358 14 2604228; fax: þ358 14 2604756. E-mail address: hkunttu@jyu.fi (H.M.T. Kunttu). 0882-4010/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.micpath.2008.10.001
bacteria, cell surface components, such as LPS and capsular material, function as virulence factors (e.g. Refs. [15–17]). Change in cell surface components often leads to change in colony morphology of a bacterium (see e.g. Refs. [16–19]). In some human and animal (also fish) pathogens, such as tubercle bacilli [20], Vibrio vulnificus [21], and Mycobacterium avium [22], different colony morphologies produced by one strain can exhibit difference in virulence. Previously, difference in colony morphology has been detected between F. columnare strains [23]. Also non-rhizoid as well as soft and non-adherent colonies have been noticed to appear among rhizoid colonies after subcultivation [24,25]. It has been shown, that spreading or rhizoid colony formation is an indication of gliding motility of Flavobacterium johnsoniae (previously Cytophaga johnsonae) [26] and Flavobacterium psychrophilum (previously Cytophaga psychrophila) [27]. Therefore, it is possible, that the loss of gliding motility appears as non-rhizoid colony morphology also in F. columnare. Different colony morphologies have also been found in Flavobacterium succinicans (previously Cytophaga succinicans), a relative to F. columnare [28]. However, the pathogenesis of colony morphology variants has not been studied in flavobacteria. We found that four different colony morphologies are formed among F. columnare strains on Shieh agar plates in laboratory cultivations, and one strain can form one or two morphology variants. Because of severity of repeated infections at fish farms and lack of knowledge of virulence mechanisms in F. columnare, we consider it important to study further these colony types. This is necessary both for developing exact diagnostic tools for columnaris infections and also to find out whether there is a connection
22
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
between virulence and colony morphology. This is why molecular characteristics and adhesion properties as well as fish mortality caused by different colony morphologies were studied. 2. Results Four different colony types (Fig. 1) were formed on Shieh agar plates in laboratory conditions by the eight F. columnare strains tested (Table 1). Colony type 1 was rhizoid and flat with yellow centre. Colony type 2 was hard, more orange in color, non-rhizoid or only slightly rhizoid, and had irregular edges and convex growth form. Colony type 3 had round edges, and smooth, yellowish appearance. Type 4 colonies were white or light yellow, smooth and spreading on the agar with irregular shape. Originally, F. columnare strains formed colony types 1–4 (Table 1), each strain producing only one colony type. In further plate cultivations other colony types started to form among original types in a following manner: type 1 / type 2, type 2 / type 4, and type 3 / type 4, meaning that among type 1 colonies, type 2 colonies started to appear, etc. These other colony types all formed in the cultures of same age and, once formed, remained the same from generation to generation in laboratory cultivations. There was no colony type change in the opposite direction meaning, that no type 3, 2 or 1 colonies were formed among type 4 colonies, no type 2 or 1 colonies among type 3 colonies, and no type 1 colonies among type 2 colonies. Colony type did not change in broth cultivations. In older cultures (more
than 3 days after plate cultivation), however, on the edges of some types 2 and 3 colonies growth resembling type 4 started to appear. There were no contaminations of other bacteria or between F. columnare strains in the cultures. Colony type 4 existed only in the genetic groups with low or intermediate virulence, whereas colony type 3 existed only in high virulence strain. On AO agar, formation of colony morphologies differed from that on Shieh agar to some extent: The growth form of bacteria was more spreading on AO agar than on Shieh agar. Moreover, the older cultures on AO agar started to get transparent appearance, which did not occur on Shieh agar. The growth form of all bacteria of all genetic groups resembled colony type 4 formed on Shieh agar, but had more spreading colonies with slightly rhizoid edges. The exceptions were genetic groups B and F, which formed types 3 and 4 (formed on Shieh agar) resembling colonies also on AO agar, respectively. Among rhizoid type 4 colonies, also types 1 and 2 (formed on Shieh agar) resembling colonies were formed by bacteria of genetic groups D, E and G. None of the colony types 1–4 formed on AO agar had exactly the same appearance as the corresponding colony types 1–4 formed on Shieh agar, but could be categorized into these groups. However, because of the spreading growth form of bacteria on AO agar, the different colony types grew rather stuck on each other than as separate colonies. On AO agar, the colony types 1, 2 and 4 formed originally on Shieh agar grew as rhizoid type 4 colonies. Types 1 and 2 (formed on Shieh agar) resembling colonies were also detected among D2 (colony type 2 formed by strain D), E1, E2 and G2. There were no differences in ARISA [10], AFLP or whole cell protein SDS-PAGE (Fig. 2) profiles between different colony types formed by one F. columnare strain. However, there were differences in adhesion capacities on polystyrene (Kruskal–Wallis test, overall comparison between strains A–H forming colony types 1–4: c2 ¼ 66,947, df ¼ 12, P < 0.001) (Fig. 3a): D2 and D4, E1 and E2, G2 and G4, as well as H2 and H4 differed significantly from each other. Similarly, on polystyrene, there was a significant difference between adhesion capacities of colony types (one-way ANOVA: F ¼ 36.679, df ¼ 4, P < 0.001) (Fig. 3b); colony type 2 was more adherent than other colony types, but also the adhesion of colony type 1 differed significantly from the other colony types. The growth rate of bacteria did not affect the adhesion capacities. On agar, colony type 2 was the most adherent followed by types 1, 3 and 4. This result was achieved by the experience when handling the colonies with inoculation loop. Adhesion capacities on gill tissue differed significantly between colony types among the strains (one-way ANOVA: F ¼ 3,306, df ¼ 6, P ¼ 0.031) (Fig. 3c). G4 was the most adherent and differed significantly from all the other colony types tested, except from E2. A significant difference between adhesion capacities of E2 and H4 was also detected. In challenge experiment, there was a significant difference between mortalities caused by the colony types (Kaplan–Meier Table 1 Colony morphology types formed on Shieh agar in laboratory conditions by genetically grouped Flavobacterium columnare strains (see Refs. [10] and [7] for more details on the strains).
Fig. 1. Representatives of different colony morphologies (1 ¼ colony type 1, 2 ¼ type 2, 3 ¼ type 3, 4 ¼ type 4) formed by Flavobacterium columnare (see also Table 1) on Shieh agar plate cultivation in laboratory. Type 1 is rhizoid and flat with yellow centre. Type 2 is hard, more orange in color, non-rhizoid or only slightly rhizoid, and has irregular edges and convex growth form. Type 3 has round edges, and smooth, yellowish appearance. Type 4 colonies are white or light yellow, smooth and spreading on the agar with irregular shape.
Genetic group
Virulence of the strain
Original colony type
Colony type formed in subcultivations
A B C D E F G H
Low High High Low High Low Intermediate Low
1 3 1 2 1a 4 2a 2a
4 2a 4a 4a
a Colony type was used in the virulence experiment using rainbow trout (Oncorhynchus mykiss, Walbaum) fingerlings.
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
M
A1
B3
C1 D2 D4
E1
E2
F4
G2 G4
H2
H4
M
a 0,080
250 130
72 55
**
0,070
OD (595 nm)
95
23
0,060
**
**
0,050
**
0,040 0,030 0,020 0,010 0,000 B3 C1 E1 E2
High virulence
36
G2 G4
A1 D2 D4 F4 H2 H4
Intermediate virulence
Low virulence
E. coli
Genetic group and colony type 28
b
0,1 0,09
* **
survival analysis: c2 ¼ 187.483, df ¼ 6, P < 0.001) (Fig. 4). Interestingly, the most polystyrene-adherent colony type 2 did not produce the highest mortality. The cumulative mortality caused by E1 was the highest, 60%, and statistically different from mortalities caused by the other colony types tested (E2, G2, G4, H2 and H4) (Table 2). The symptoms of columnaris disease (skin and/or gill lesions) were detectable in all fish infected with E1, but present approximately only in half of the fish infected with the other colony types. F. columnare could be isolated only from fish infected with colony type E1. Interestingly, from primary cultivations made from these fish, several colony types per one fish could be isolated: in addition to colony type 1, the isolated bacteria were also of colony types 2 and 4. 3. Discussion This is the first time when spontaneously formed colony morphologies among F. columnare strains are studied to find connection between colony types, adherence and virulence. Colony types among each F. columnare strain showed no genomic differences in ARISA, AFLP, or whole cell protein SDS-PAGE profiles. The four colony types formed on Shieh agar by F. columnare strains differed significantly in their adhesion capacities on polystyrene and agar; type 2 was the most adherent followed by types 1, 3 and 4. Virulence, however, was not related to adherence capacity: strongly polystyrene- and agar-adherent type 2 bacteria (E2, G2 and H2) were not more virulent than rhizoid type 1 (E1) or poorly adherent type 4 (G4 and H4) bacteria. F. columnare could be isolated only from fish dead in E1 infections, which may be due to difficulties in isolation of F. columnare at preclinical stages of infection, or possible presence of competitive bacteria [29]. Our results suggest that strong adhesion is not essential for virulence of F. columnare. In addition, adhesion to gill tissue was not related either to virulence or adherence on polystyrene, indicating that same molecules are not used to attach on different surfaces, and that virulence mechanisms other than adhesion may be important in pathogenesis of F. columnare. Formation of different colony types occurred also when F. columnare was grown on AO agar, but colonies were not exactly similar to and did not grow as clear separate colonies as the colonies formed on Shieh agar. Differential colony formation on Shieh and AO agar may have been caused by different nutrition composition between these two media. Shieh medium is
*
0,07
* **
0,06
* **
* **
0,05 0,04 0,03 0,02 0,01 0
1
2
3
4
Colony type
c
500
*
450
*
400
CFU ml-1 x 106
Fig. 2. Whole cell protein SDS-PAGE profiles of Flavobacterium columnare strains A–H forming colony types 1–4 (see Table 1). In the first and the last lane are the molecular mass markers (kDa).
OD (595 nm)
0,08
**
* *
350
*
300 250 200 150 100 50 0 C1
E1
E2
High virulence
G2
G4
Intermediate virulence
H2
H4
Low virulence
Genetic group and colony type Fig. 3. (a). Adherence (mean SE) on polystyrene of Flavobacterium columnare strains A–H forming colony types 1–4 (see Table 1), and Escherichia coli. Statistical differences (Kruskal–Wallis test, pairwise comparisons) between different colony types formed by one strain: * ¼ P < 0.05, ** ¼ P < 0.01, *** ¼ P < 0.001. (b). Adherence (mean SE) on polystyrene of Flavobacterium columnare strains (A–H) representing the same colony morphology type (1–4) (see Table 1). Statistical differences (one-way ANOVA, pairwise comparisons, Post Hoc LSD test) between different colony types: * ¼ P < 0.05, ** ¼ P < 0.01, *** ¼ P < 0.001. (c). Adherence (mean SE) on gill tissue (rainbow trout (Oncorhynchus mykiss, Walbaum) of Flavobacterium columnare strains C, E, G and H forming colony types 1–4 (see Table 1). Statistical differences (one-way ANOVA, pairwise comparisons, Post Hoc LSD test) between the strains: * ¼ P < 0.05, ** ¼ P < 0.01, *** ¼ P < 0.001).
specific for F. columnare [5] and might thus allow broader variation of colonial morphology than AO medium. Change in colony morphology is not a rare phenomenon; it occurs in many bacterial species pathogenic to humans and animals. For example, colony morphology variation between translucent and opaque colony types is observed both in Vibrio parahaemolyticus [30] and V. vulnificus [21], whereas M. avium [22]
24
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
100 E1 E2 G2 G4 H2 H4 Control
Cumulative mortality (%)
90 80 70 60 50 40 30 20 10 0 0
24
48
72
96
120
144
168
192
Time (hours after exposure) Fig. 4. Average cumulative mortality of rainbow trout (Oncorhynchus mykiss, Walbaum) fingerlings challenged with equal number of bacteria representing different morphology types (types 1, 2 and 4, see Table 1) formed by Flavobacterium columnare genetic groups E (high virulence), G (intermediate virulence) and H (low virulence).
represents smooth and rough colony types. Colony morphology change is sometimes caused by a phenomenon known as phase variation (reviewed in Refs. [31,32]). Phase variation is often thought to be a random event occurring in a bacterial population as a consequence of heritable genomic changes which can be seen as subpopulations of different phenotypes [33]. Most of the changes are caused by more or less radical genomic rearrangements, like DNA inversions or recombinational deletions, but some minor DNA modifications, like differential DNA methylation, are also known to lead to phase variation. However, growing amount of evidence suggests that phase variation is not always a random event, but a response to an environmental change and caused by change in gene expression [31]. Phase variation allows bacteria to adapt the surrounding environment and is caused by a reversible on/off switch in the expression of one or more genes [32]. Switch can be triggered by intra- or extracellular activators or inhibitors, such as temperature [34], or signaling molecules received from the environment [32]. These molecules can be related to e.g. bacterial quorum sensing. Quorum sensing is a chemical signaling system which is used by bacterial populations to control variety of functions via regulation of gene expression [35]. Besides chemicals produced by bacteria, molecules received from a host can also affect behavior of bacterial populations [36,37]. Production of non-typical colony morphologies after subcultivation has been previously detected also in F. columnare [25]. In another study, non-adherent F. columnare colony variant was obtained by b-lactam selection [24]. In these studies, no genetic methods were used to find out whether different colony formation is caused by changes at the DNA level. Studying the properties of different colony types can provide information about virulence and
Table 2 Difference between colony morphology types 1, 2 and 4 formed by Flavobacterium columnare genetic groups E, G and H (see also Table 1) in virulence on rainbow trout (Oncorhynchus mykiss, Walbaum) fingerlings (Kaplan–Meier survival analysis, pairwise comparisons, Mantel Cox). Genetic group and colony type
E2 G2 G4 H2 H4 a
Difference in virulence: P-values E1
E2
G2
G4
H2
0.000a 0.000a 0.000a 0.000a 0.000a
0.587 0.787 0.002a 0.070
0.778 0.007a 0.197
0.004a 0.119
0.081
Statistically significant difference in virulence.
transmission strategies of F. columnare. It is possible that change in colony type of F. columnare is caused by phase variation affecting the expression or modification of some bacterial surface molecule. However, because no genomic changes were detectable by ARISA, AFLP or whole cell protein SDS-PAGE, the changes may be only minor or may take place at the gene expression level. In human pathogens E. coli and Salmonella enterica serotype Typhimurium phase variation occurs as a consequence of different DNA methylation, which does not lead to change in DNA sequence [32]. In V. parahaemolyticus genetic alterations in cell type controlling transcription factor coding opaR gene may lead to formation of different colony types [30]. Although phase variation is usually considered as a reversible phenomenon, also irreversible changes in bacterium’s phenotype can be caused. For example, irreversible phase variation occurs in a human pathogen V. vulnificus [18]. Also in the present study, the change in colony type was found to be irreversible. Possible phase variation occurring in F. columnare may be a consequence of a change in gene expression and a response to a change in growth environment. It is known, for example, that phase variation occurs during starvation in V. vulnificus [38] and Neisseria gonorrhoeae [39]. The effect of change in nutritional conditions of the growth environment on F. columnare phenotype could be seen also in this study as difference between colony morphology formation on Shieh and AO agars. In addition to changes in growth environment, also quorum sensing has a central role in phase variation and biofilm development in some bacteria, for example in V. parahaemolyticus [19]. Biofilm production properties are needed in the beginning of V. cholerae infection, but suppressed later by quorum sensing [40]. Quorum sensing also controls virulence gene expression in V. cholerae [41]. Quorum sensing has not been studied in F. columnare, and in F. psychrophilum no evidence of quorum sensing was detected, when acylated homoserine lactone (AHL)-based quorum sensing signaling system was studied [42]. However, among AHLbased signaling also several other quorum sensing systems are known and the possible existence of quorum sensing system in F. columnare cannot be excluded. Isolation of colony types 1, 2 and 4 from fish infected with E1 indicates that different colony types may have different roles in the infection cycle, but whether they are formed by phase variation as a response to nutritional factors or quorum sensing can only be speculated and needs more studies. However, it is known that, especially in an animal host, the environment can drive the selection against a particular phenotype of a bacterium, which can result in a different phenotype for cells cultured in the laboratory and those isolated from the host [32]. Indeed, phase variation is considered as one mechanism of pathogenic bacteria to avoid immune defense of a host [31]. Our results show, that there is no clear relation between adhesion capacity and virulence of F. columnare strains tested in this study. We have previously tested adhesion capacity of F. columnare also on crude mucus-coated slides [7], and similarly no correlation to virulence was detected. These results contradict the previous study by Bader et al. [24]. In their immersion challenge experiment the smooth colony forming, less virulent, F. columnare mutant was significantly less adherent both on skin and gill tissue of channel catfish (Ictalurus punctatus) than wild type strain, which formed hard colonies with finger-like projections and which may have resembled our colony type 1. However, in their study, Bader et al. did not detect colony morphologies which would have been adherent and hard with no finger-like projections, i.e. similar to our colony type 2. This may explain the difference in results compared to our study: the virulent strain in that study resembling our high virulence colony type 1 was also the most adherent and Bader et al. did not detect highly adherent type 2 colonies, which produced less mortality in our study. There is also a contradiction with the study by Decostere et al. [9], in which study one highly virulent
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
F. columnare strain was more adherent on the gill tissue of black mollies (Poecilia sphenops) than a less virulent strain. Decostere et al. [9] did not provide any further information on the bacterial colonies, so comparison between their and our study is more difficult to make. In our study, the results from eight genetically different strains, representing high, intermediate and low virulence strains, suggest that adherence capacity is not the main factor defining virulence in F. columnare. Further, the adhesion factors of F. columnare in polystyrene adhesion are different from those in gill tissue adhesion. In summary, we suggest, that strong adhesion capacity is not the defining virulence factor of F. columnare. Thus, some other virulence factors, such as chondroitin AC lyase activity [7] or features not yet found, are needed for pathogenesis in this bacterium. In addition, rhizoid colony morphology may also be related to virulence. This has already been shown in F. psychrophilum, when loss of gliding motility led to decreased virulence [43]. Moreover, in addition to F. psychrophilum [27], relation between rhizoid colony formation and gliding motility has been shown on F. johnsoniae [26,44]. However, our observation is based only on one F. columnare strain (E1), but the virulence experiment suggests similar gliding motility–virulence relationship as found in F. psychrophilum. To confirm this, more studies on the effects of colony morphology on bacterial virulence are needed. 4. Materials and methods 4.1. Bacterial strains and isolation of colony types F. columnare type strain F (NCIMB 2248T) and Finnish strains A–E and G–H isolated from diseased fish at different fish farms (see Ref. [10]) and exhibiting different degrees of virulence [7], were stored in 80 C in 500 ml stocks containing 10% fetal calf serum (FCS, Gibco, BRL Co.) and 10% glycerol, thawed and inoculated to 10 ml of Shieh broth [5] (broth I). Cultures were grown at room temperature (RT) under constant shaking (120 rpm) to late log phase (23 h), after which the cultures were plated on Shieh agar (plate I). Plate cultures were grown for 3 days, different colony types isolated with a loop and inoculated to 10 ml of Shieh broth (broth II). Cultures were plated as previously (plate II), grown for 4 days, different colony types isolated and grown in Shieh broth (broth III) and plated (plate III) as previously. From broth III (broth II in case of strain F), stocks were made and stored in 80 C. Plate I cultures were photographed using stereo microscope on days 2, 3 and 6, plate II cultures on days 2 and 4, and plate III cultures on days 2, 3, 4, 7, 9, 10, 17 and 22 after cultivation. For strain F only plate I and II cultures were made, and they were photographed on days 4, 7, 9, 10, 17 and 22 after cultivation. Several consecutive cultivations were made in order to check whether the colony types retained their morphology. For subsequent analyses, bacterial stocks made from broth III (broth II in case of strain F) were thawed and grown in 10 ml of Shieh broth to late log phase. Colony formation on another medium, AO broth and AO agar [45], was also tested. Experiment on AO medium was conducted as follows: F. columnare strains mentioned above were thawed, inoculated to 10 ml of AO broth and grown at RT under constant shaking (100 rpm) for 42 h, after which the cultures were plated on AO agar. Agar cultures were examined for colony formation using stereo microscope on days 2, 5, 7, 19 and 21 after cultivation. Also the colony types originally formed on Shieh agar (see Table 1) were cultivated in AO broth and on AO agar plates and plate cultures were examined as above. 4.2. Adhesion on polystyrene Optical densities (OD) of bacterial cultivations at 595 nm were measured with micro plate reader (Multiskan Plus, Labsystems,
25
Vantaa, Finland) and set to OD 0.260 in 200-ml volume, corresponding to 3.0 108 colony forming units (CFU) ml1 of bacteria. For adhesion, 100 ml of each colony type culture and Escherichia coli strain JM109 as a control was inoculated onto 96-well polystyrene plates (Immuno Plate, Maxisorp, Nunc Co., Denmark) in six replicates. Another plate with four replicate wells of each colony type was used to detect bacterial growth rate during adhesion experiment. Bacteria were allowed to adhere on the plate for 2 h at RT. Relative adherence of bacterial cells was determined by the method described by A´lvarez et al. [43]. Adhesion capacities were determined from OD values at 595 nm after crystal violet staining of adhered bacteria. ODs of growth controls at 595 nm were measured in the beginning and after the adhesion incubation. 4.3. Adhesion on gill tissue The outermost gill arches (mean weight 30 mg) were taken from 12 rainbow trout fry (mean weight 23 g) and exposed to broth cultures of four F. columnare strains (C, E, G and H) representing colony types 1, 2 and 4 (C1, E1, E2, G2, G4, H2 and H4, see Table 1). ODs of bacterial cultivations were set to 0.119 (corresponding to 1.29 108 CFU ml1). One gill arch per treatment was placed on a petri dish and 15 ml of bacterial culture was added. Shieh broth was used as a negative control. For each bacterium, three replicate dishes were made. Bacteria were allowed to adhere on gill tissue for 2 h at RT under constant shaking (50 rpm). After adhesion, gill arches were rinsed with 10 ml of sterile distilled water and homogenized in eppendorf tubes containing 300 ml phosphatebuffered saline (PBS). From homogenate 1:10 serial dilutions were made in PBS, and 100 ml of pure homogenate and each dilution were plated on Shieh agar containing tobramycin, 1 mg ml1 [4]. Plates were incubated at RT for 2 days after which plate counts were made and the adhesion capacities of the colony types determined from CFU ml1 values. 4.4. ARISA (automated ribosomal intergenic spacer analysis) Genomic DNA was extracted from colony types grown in Shieh broth using microbial DNA isolation kit (MOBIO). ARISA was conducted as previously [10] with following modifications: reaction mixtures contained 1.0 mM of each primer, 0.2 mM of dNTPs, 1 Biotools DNA polymerase buffer (Biotools), and 0.3 U of Biotools DNA polymerase (Biotools). The total reaction volume in each PCR tube was 30 ml plus 1.0 ml of template DNA, and the temperature of an initial denaturation step 95 C. 4.5. AFLP (amplified fragment length polymorphism) Genomic DNA of colony types was extracted as described above and subjected to AFLP based on Vos et al. [46]. Genomic DNA was digested successively with two restriction enzymes, EcoRI and MseI. Reaction solution (total volume 5.1 ml per sample), containing 2.0 U of EcoRI (Fermentas), 2.0 U of MseI (Fermentas), 0.5 ml of Rþ buffer (Fermentas) and 4.2 ml DNA (ca. 20 ng), was incubated first at 37 C for 1 h (for EcoRI digestion) and then at 65 C for 2 h (for MseI digestion). By heating at 95 C for 5 min and then cooling slowly, double stranded adapters for restriction sites were hybridized from single stranded EcoRI forward and EcoRI reverse (5.0 mM suspension), and MseI ad1 and MseI ad2 (50 mM suspension) adapters. These two stranded adapters were ligated at the ends of restriction fragments by incubating ligation solution (total volume 10 ml per sample), including 1.0 U of T4 DNA ligase (Fermentas), 1 T4 ligation buffer (Fermentas), 0.1 mM of EcoRI adapter, 1.0 mM of MseI adapter, 3.4 ml of sterile distilled water and 5.0 ml of digestion product, at RT for 1 h. The ligation products were used as templates in PCR, in which adapter specific primers were used: one reaction
26
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27
was done using primer pair MseI-0/EcorRI-A-IRD700 and another using MseI-0/EcorRI-ACT-IRD800, in which A and ACT stand for additional selective nucleotides and 0 for no selective nucleotide. Twenty microliters of PCR reaction mixture, containing 0.25 mM of each primer, 1.0 mM of MgCl2 (Fermentas), 0.08 mM of dNTPs, 0.2 U of Taq DNA polymerase (Fermentas), 1 Tag DNA polymerase buffer [(NH4)2SO4 added] (Fermentas), 11.2 ml of sterile distilled water, and 3.0 ml of template, was made per each sample. PCR was run according to following program: initial denaturation step at 94 C for 2 min, 13 cycles of amplification I (94 C for 30 s, 65 C for 30 s, 72 C for 2 min), 24 cycles of amplification II (94 C for 30 s, 56 C for 30 s, 72 C for 2 min), and final elongation step at 72 C for 10 min. The AFLP was run with a LI-COR 4200 automatic sequencer (LI-COR BioTech) in 6% Long Ranger denaturing polyacrylamide gel (FMC Bioproducts), and fragment length data was analyzed using IrfanView 3.97 free software.
4.6. SDS-PAGE (sodium dodecyl sulphate-polyacrylamide gel electrophoresis) of whole cell proteins Whole cell proteins were extracted from broth cultures of colony types according to protocol by Tan et al. [47] with slight modifications. Bacterial culture, 100 ml, was centrifuged at 4332 g for 10 min. The pellet was washed once with 50 ml of 10 mM Tris-HCl buffer, pH 7.6, and resuspended in 20 ml of the buffer. The extract was sonicated on ice (Branson Digital SonifierÒ, Branson Ultrasonic Corporation, Danburry, Connecticut, USA) for 1 min (1 s pulse with 2 s interval), and 500 ml of sonicate was suspended in an equal volume of 2 electrophoresis treatment buffer and stored in 20 C until analyzed. Prior electrophoresis, the samples were heated at 100 C for 2 min and placed on ice. Discontinuous polyacrylamide gel, 16 cm long and 1.5 mm thick, was used. Samples (50 ml) were pipetted into wells, and electrophoresis was run first with 120 V for 1 h, then with 180 V for 4 h, and finally with 200 V for 1 h. The protein patterns were visualized by staining with PageBlueÔ Protein Staining Solution (Fermentas).
4.7. Challenge experiment Three F. columnare strains (E, G and H) representing colony types 1, 2 and 4 (E1, E2, G2, G4, H2 and H4, see Table 1), were tested for their virulence in rainbow trout (Oncorhynchus mykiss, Walbaum) fingerlings. Strain E has previously been defined as high virulence, G as intermediate virulence, and H as low virulence strain [7]. These strains were chosen because in the present study it was not possible to carry out an experimental set up with all colony types formed by all genetic groups. Fish, (mean weight 6.0 g) originated from a stock of a fish farm in Central Finland, and were acclimated in laboratory conditions for 3 weeks before the experiment. Bacterial infection of fish was carried out under constant aeration in 10-l plastic aquaria, 25 fish in each aquarium, with well water containing 6.5 106 CFU ml1 of bacteria. Each bacterial infection was done in triplicate, a total of 75 fish per treatment. Fish treated with Shieh broth without bacterial inoculate served as negative control. The water temperature was 25 C and infection time 2 h. After infection, the fish were dipped in pure well water and transferred at random into clean, aerated, 10-l plastic flow troughaquaria (T ¼ 25 C). Fish were monitored for 8 days by 12-h intervals for signs of disease and mortality. Bacterial cultivations from skin and gill lesions of up to ten dead fish per treatment were made on Shieh agar. At the end of the experiment cultivations were made from the trunk kidney of four fish per treatment. Cumulative mortality of fish was used as an indicator of bacterial virulence.
4.8. Statistics Differences between adhesion capacities of F. columnare strains forming colony types 1–4 were analyzed using SPSS 14.0 software. In the polystyrene adhesion experiment, comparisons between OD values of each strain and their colony types were made using nonparametric Kruskal–Wallis test. Means of OD values of each colony type were calculated and compared using one-way ANOVA. CFU counts of gill tissue adhesion were analyzed using one-way ANOVA. Data from the bacterial challenge experiment was analyzed using Kaplan–Meier survival analysis. P-values below 0.05 were considered significant.
Acknowledgements The study was supported by a grant from Academy of Finland (# 110754). The authors would like to thank Ellen and Artturi Nyysso¨nen foundation for funding H. M. T. Kunttu and Graduate School of Biological Interactions for other support. We thank also F. Teixeira and M. Martins for technical help and would also acknowledge Dr. P. Rintama¨ki (University of Oulu) and Dr. V. Hirvela¨-Koski (Finnish Food Safety Authority, Oulu) for kindly providing some of the bacterial isolates used in this study.
References [1] Revetta RP, Rodgers MR, Kinkle BK. Isolation and identification of freshwater bacteria antagonistic to Giardia intestinalis cysts. J Water Health 2005;3:83–8. [2] Rickard AH, McBain AJ, Ledder RG, Handley PS, Gilbert P. Coaggregation between freshwater bacteria within biofilm and planktonic communities. FEMS Microbiol Lett 2003;220:133–40. [3] Wagner BA, Wise DJ, Khoo LH, Terhune JS. The epidemiology of bacterial diseases in food-size channel catfish. J Aquat Anim Health 2002;14:263–72. [4] Decostere A, Haesebrouck F, Devriese LA. Shieh medium supplemented with tobramycin for selective isolation of Flavobacterium columnare (Flexibacter columnaris) from diseased fish. J Clin Microbiol 1997;35:322–4. [5] Shieh HS. Studies on the nutrition of a fish pathogen, Flexibacter columnaris. Microbios Lett 1980;13:129–33. [6] Shoemaker CA, Olivares-Fuster O, Arias CR, Klesius PH. Flavobacterium columnare genomovar influences mortality in channel catfish (Ictalurus punctatus). Vet Microbiol 2007;127:353–9. [7] Suomalainen L-R, Tiirola M, Valtonen ET. Chondroitin AC lyase activity is related to virulence of fish pathogenic Flavobacterium columnare. J Fish Dis 2006;29:757–63. [8] Thomas-Jinu S, Goodwin AE. Morphological and genetic characteristics of Flavobacterium columnare isolates: correlations with virulence in fish. J Fish Dis 2004;27:29–36. [9] Decostere A, Haesebrouck F, Charlier G, Ducatelle R. The association of Flavobacterium columnare strains of high and low virulence with gill tissue of black mollies (Poecilia sphenops). Vet Microbiol 1999;67:287–98. [10] Suomalainen L-R, Kunttu H, Valtonen ET, Hirvela¨-Koski V, Tiirola M. Molecular diversity and growth features of Flavobacterium columnare strains isolated in Finland. Dis Aquat Org 2006;70:55–61. [11] Bertolini JM, Rohovec JS. Electrophoretic detection of proteases from different Flexibacter columnaris strains and assessment of their variability. Dis Aquat Org 1992;12:121–8. [12] Newton JC, Wood TM, Hartley MM. Isolation and partial characterization of extracellular proteases produced by isolates of Flavobacterium columnare derived from channel catfish. J Aquat Anim Health 1997;9:75–85. [13] Xie HX, Nie P, Sun BJ. Characterization of two membrane-associated protease genes obtained from screening out-membrane protein genes of Flavobacterium columnare G4. J Fish Dis 2004;27:719–29. [14] Zhang Y, Arias CR, Shoemaker CA, Klesius PH. Comparison of lipopolysaccharide and protein profiles between Flavobacterium columnare strains from different genomovars. J Fish Dis 2006;29:657–63. [15] Amaro C, Fouz B, Biosca EG, Marco-Noales E, Collado R. The lipopolysaccharide O side chain of Vibrio vulnificus serogroup E is a virulence determinant for eels. Infect Immun 1997;65:2475–9. [16] Arias CR, Shoemaker CA, Evans JJ, Klesius PH. A comparative study of Edwardsiella ictaluri parent (EILO) and E. ictaluri rifampicin-mutant (RE-33) isolates using lipopolysaccharides, outer membrane proteins, fatty acids, Biolog, API 20E and genomic analyses. J Fish Dis 2003;26:415–21. [17] Yoshida S, Ogawa M, Mizuguchi Y. Relation of capsular materials and colony opacity to virulence of Vibrio vulnificus. Infect Immun 1985;47:446–51. [18] Hilton T, Rosche T, Froelich B, Smith B, Oliver J. Capsular polysaccharide phase variation in Vibrio vulnificus. Appl Environ Microbiol 2006;72:6986–93.
H.M.T. Kunttu et al. / Microbial Pathogenesis 46 (2009) 21–27 [19] Enos-Berlage JL, Guvener ZT, Keenan CE, McCarter LL. Genetic determinants of biofilm development of opaque and translucent Vibrio parahaemolyticus. Mol Microbiol 2005;55:1160–82. [20] Smithburn KC. The colony morphology of tubercle bacilli. III. The relation between virulence and colony form. J Exp Med 1935;62:645–63. [21] Simpson LM, White VK, Zane SF, Oliver JD. Correlation between virulence and colony morphology in Vibrio vulnificus. Infect Immun 1987;55:269–72. [22] Kansal RG, Gomez-Flores R, Mehta RT. Change in colony morphology influences the virulence as well as the biochemical properties of the Mycobacterium avium complex. Microb Pathog 1998;25:203–14. [23] Song YL, Fryer JL, Rohovec JS. Comparison of gliding bacteria isolated from fish in North America and other areas of the Pacific Rim. Fish Pathol 1988;23:197–202. [24] Bader JA, Shoemaker CA, Klesius PH. Production, characterization and evaluation of virulence of an adhesion defective mutant of Flavobacterium columnare produced by b-lactam selection. Lett Appl Microbiol 2005;40:123–7. [25] Bernardet JF.´Flexibacter columnaris´: first description in France and comparison with bacterial strains from other origins. Dis Aquat Org 1989;6:37–44. [26] Chang LE, Pate JL, Betzig RJ. Isolation and characterization of nonspreading mutants of the gliding bacterium Cytophaga johnsonae. J Bacteriol 1984;159:26–35. [27] Bernardet JF, Kerouault B. Phenotypic and genomic studies of ‘‘Cytophaga psychrophila’’ isolated from diseased rainbow trout (Oncorhynchus mykiss) in France. Appl Environ Microbiol 1989l;55:1796–800. [28] Anderson RL, Ordal EJ. Cytophaga succinicans sp. n., a factaltatively anaerobic, aquatic myxobacterium. J Bacteriol 1961;81:130–8. [29] Suomainen L-R, Tiirola MA, Valtonen ET. Effect of Pseudomonas sp. MT5 baths on Flavobacterium columnare infection of rainbow trout and on microbial diversity on fish skin and gills. Dis Aquat Org 2005;63:61–8. [30] McCarter LL. OpaR, a homolog of Vibrio harveyi LuxR, controls opacity of Vibrio parahaemolyticus. J Bacteriol 1998;180:3166–73. [31] van der Woude MW. Re-examining the role and random nature of phase variation. FEMS Microbiol Lett 2006;254:190–7. [32] van der Woude MW, Baumler AJ. Phase and antigenic variation in bacteria. Clin Microbiol Rev 2004;17:581–611. [33] Henderson IR, Owen P, Nataro JP. Molecular switches – the ON and OFF of bacterial phase variation. Mol Microbiol 1999;33:919–32.
27
[34] Ratiner YA. Temperature-dependent flagellar antigen phase variation in Escherichia coli. Res Microbiol 1999;150:457–63. [35] Bassler BL. How bacteria talk to each other: regulation of gene expression by quorum sensing. Curr Opin Microbiol 1999;2:582–7. [36] Daniels R, Vanderleyden J, Michiels J. Quorum sensing and swarming migration in bacteria. FEMS Microbiol Rev 2004;28:261–89. [37] Gonzalez JE, Keshavan ND. Messing with bacterial quorum sensing. Microbiol Mol Biol Rev 2006;70:859–75. [38] Chatzidaki-Livanis M, Jones MK, Wright AC. Genetic variation in the Vibrio vulnificus group 1 capsular polysaccharide operon. J Bacteriol 2006; 188:1987–98. [39] Stohl AE, Seifert HS. The recX gene potentiates homologous recombination in Neisseria gonorrhoeae. Mol Microbiol 2001;40:1301–10. [40] Higgins DA, Pomianek ME, Kraml CM, Taylor RK, Semmelhack MF, Bassler BL. The major Vibrio cholerae autoinducer and its role in virulence factor production. Nature 2007;450:883–6. [41] Zhu J, Miller MB, Vance RE, Dziejman M, Bassler BL, Mekalanos JJ. Quorumsensing regulators control virulence gene expression in Vibrio cholerae. Proc Natl Acad Sci U S A 2002;99:3129–34. [42] Bruhn JB, Dalsgaard I, Nielsen KF, Buchholtz C, Larsen JL, Gram L. Quorum sensing signal molecules (acylated homoserine lactones) in gram-negative fish pathogenic bacteria. Dis Aquat Organ 2005;65:43–52. [43] Alvarez B, Secades P, Prieto M, McBride MJ, Guijarro JA. A mutation in Flavobacterium psychrophilum tlpB inhibits gliding motility and induces biofilm formation. Appl Environ Microbiol 2006;72:4044–53. [44] Hunnicutt DW, McBride MJ. Cloning and characterization of the Flavobacterium johnsoniae gliding motility genes gldD and gldE. J Bacteriol 2001;183:4167–75. [45] Anacker RL, Ordal EJ. Study of a bacteriophage infecting myxobacterium Chondrococcus columnaris. J Bacteriol 1955;70:738–41. [46] Vos P, Hogers R, Bleeker M, Reijans M, Van De Lee T, Hornes M, et al. AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res 1995; 23:4407–14. [47] Tan ZY, Xu XD, Wang ET, Gao JL, Martinez-Romero E, Chen WX. Phylogenetic and genetic relationships of Mesorhizobium tianshanense and related rhizobia. Int J Syst Bacteriol 1997;47:874–9.